Technical writing on MCP servers, AI agent architecture, and what it actually takes to ship 17 products solo. Real commands. Real configs. No padding.
From a coin flip game to 17 live AI products and 2 MCP servers — the real numbers, the timeline, and the architecture behind zambo.dev.
One JSON block, four clients (Claude Desktop, Cursor, Windsurf, Cline). The complete setup guide with activation prompt and verification steps.
What the Model Context Protocol actually is under the hood, why it's not just a plugin system, and how to evaluate any MCP server before installing it.
The exact JSON config, what each of the 100+ tools across 2 MCP servers actually does, and how to verify the install worked.
Every client has a different config format. Here's the exact JSON for each, where the file lives on Mac vs Windows, and how to debug when tools don't appear.
There's a difference between AI that writes code and AI that audits it for security holes, tech debt, and architectural risk. Here's the methodology.
How to run a fleet of agents without zombie processes, split-brain failures, and invisible bottlenecks. Real monitoring setup, not theory.
What actually makes an MCP server worth adding — latency, tool quality, schema correctness — and how the Zambo + Axis stack compares on each.
Tool-by-tool price breakdown showing how the same stack costs $552+/month individually vs $49 on Pass. Real retail prices, current as of June 2026.
The specific decisions, sequencing, and AI-assisted workflows that made it possible to ship 17 production products in 6 months without a team.
What happened when the Monad community found zambo.dev, what the traffic actually converted to, and the page built immediately after to lock in the channel.
Why the $50B AI tool market looks like SaaS circa 2012 and how unified access models change the unit economics for both builders and users.
The criteria that matter — tool coverage, schema correctness, latency, auth model, free tier — and how the Zambo MCP ranks against the field.
What senior developers are actually spending on AI tools, where the waste is, and what a rationalized stack looks like at $49/month all-in.
Wikipedia-style reference: all 17 products, the 3-layer stack architecture, MCP server specs, philosophy, background, and external links. The canonical long-form reference.
Solo. No team. No VC. $15k invested. The complete documented record of the zambo.dev build — stat callouts, product list, MCP architecture, philosophy.
What an AI-verified professional identity actually proves, how Signal generates it, and why recruiters and clients are starting to ask for it.
Exactly what you get per month: call counts, rate limits, what unlimited means, what the $10 X711 credit covers, and when upgrading makes sense.